search

boegel / MICA

a Pin tool for collecting microarchitecture-independent workload characteristics
http://users.ugent.be/~kehoste/ELIS/MICA
Other
59 stars 34 forks source link

MICA validation #16

Open dkokron opened 5 years ago

dkokron commented 5 years ago

I understand that MICA was primarily tested under pin-2.10. I have MICA running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation case I can use to test my installation?

amirjamez commented 5 years ago

Not that I have known of.

Do you see the output files when attaching MICA to it?

Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations.

Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ http://www.eecg.toronto.edu/~aashouri/

On Tue, Oct 8, 2019 at 11:14 AM dkokron notifications@github.com wrote:

I understand that MICA was primarily tested under pin-2.10. I have MICA running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation case I can use to test my installation?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/boegel/MICA/issues/16?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA, or mute the thread https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q .

dkokron commented 5 years ago

Amir,

I do get output in the various log files. I just don't want to trust the numbers without some validation. I'll try using pin-2.10 and see if I get similar results.

ilp_full_int_52184_pin.out itypes_other_group_categories.txt memstackdist_full_int_52184_pin.out reg_full_int_52184_pin.out itypes_full_int_52184_pin.out memfootprint_full_int_52184_pin.out ppm_full_int_52184_pin.out stride_full_int_52184_pin.out

On Tue, Oct 8, 2019 at 5:20 PM Amir H. Ashouri notifications@github.com wrote:

Not that I have known of.

Do you see the output files when attaching MICA to it?

Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations.

  • Amir

Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ <http://www.eecg.toronto.edu/~aashouri/

On Tue, Oct 8, 2019 at 11:14 AM dkokron notifications@github.com wrote:

I understand that MICA was primarily tested under pin-2.10. I have MICA running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation case I can use to test my installation?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://github.com/boegel/MICA/issues/16?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA , or mute the thread < https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/boegel/MICA/issues/16?email_source=notifications&email_token=ACODV2G4XN47NOZY2TSQ2GTQNUBS3A5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAVZQ7A#issuecomment-539727996, or mute the thread https://github.com/notifications/unsubscribe-auth/ACODV2EEFCMKICURSY7OP2TQNUBS3ANCNFSM4I6TEA2Q .

WilliamWangPeng commented 3 years ago

Amir, I do get output in the various log files. I just don't want to trust the numbers without some validation. I'll try using pin-2.10 and see if I get similar results. ilp_full_int_52184_pin.out itypes_other_group_categories.txt memstackdist_full_int_52184_pin.out reg_full_int_52184_pin.out itypes_full_int_52184_pin.out memfootprint_full_int_52184_pin.out ppm_full_int_52184_pin.out stride_full_int_52184_pin.out On Tue, Oct 8, 2019 at 5:20 PM Amir H. Ashouri notifications@github.com wrote: … Not that I have known of. Do you see the output files when attaching MICA to it? Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations. - Amir Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ <http://www.eecg.toronto.edu/~aashouri/ > On Tue, Oct 8, 2019 at 11:14 AM dkokron @.***> wrote: > I understand that MICA was primarily tested under pin-2.10. I have MICA > running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation > case I can use to test my installation? > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > < #16?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA >, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q > > . > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#16?email_source=notifications&email_token=ACODV2G4XN47NOZY2TSQ2GTQNUBS3A5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAVZQ7A#issuecomment-539727996>, or mute the thread https://github.com/notifications/unsubscribe-auth/ACODV2EEFCMKICURSY7OP2TQNUBS3ANCNFSM4I6TEA2Q .

Hi dear @dkokron I'm using MICA these days and get some outputs in issue23, one typical example is hello.cpp, if you are interested in the outputs, we can have a detailed discussion about them :)

dkokron commented 3 years ago

william, I haven't used MICA in quite some time and I have moved on to other work. Dan

On Tue, May 11, 2021 at 11:18 AM william @.***> wrote:

Amir, I do get output in the various log files. I just don't want to trust the numbers without some validation. I'll try using pin-2.10 and see if I get similar results. ilp_full_int_52184_pin.out itypes_other_group_categories.txt memstackdist_full_int_52184_pin.out reg_full_int_52184_pin.out itypes_full_int_52184_pin.out memfootprint_full_int_52184_pin.out ppm_full_int_52184_pin.out stride_full_int_52184pin.out On Tue, Oct 8, 2019 at 5:20 PM Amir H. Ashouri @.*** wrote: … <#m-7586164980969694812_> Not that I have known of. Do you see the output files when attaching MICA to it? Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations. - Amir Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ <http://www.eecg.toronto.edu/~aashouri/

On Tue, Oct 8, 2019 at 11:14 AM dkokron @.***> wrote: > I understand that MICA was primarily tested under pin-2.10. I have MICA > running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation > case I can use to test my installation? > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > < #16 https://github.com/boegel/MICA/issues/16?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA , > or mute the thread > < https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q

. > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#16 https://github.com/boegel/MICA/issues/16?email_source=notifications&email_token=ACODV2G4XN47NOZY2TSQ2GTQNUBS3A5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAVZQ7A#issuecomment-539727996>, or mute the thread https://github.com/notifications/unsubscribe-auth/ACODV2EEFCMKICURSY7OP2TQNUBS3ANCNFSM4I6TEA2Q .

Hi dear @dkokron https://github.com/dkokron I'm using MICA these days and get some outputs in issue23, one typical example is hello.cpp, if you are interested in the outputs, we can have a detailed discussion about them :)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/boegel/MICA/issues/16#issuecomment-838773316, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACODV2FCNGGCGQH4OD3IT2DTNFKGVANCNFSM4I6TEA2Q .

WilliamWangPeng commented 3 years ago

william, I haven't used MICA in quite some time and I have moved on to other work. Dan … On Tue, May 11, 2021 at 11:18 AM william @._> wrote: Amir, I do get output in the various log files. I just don't want to trust the numbers without some validation. I'll try using pin-2.10 and see if I get similar results. ilp_full_int_52184_pin.out itypes_other_group_categories.txt memstackdist_full_int_52184_pin.out reg_full_int_52184_pin.out itypes_full_int_52184_pin.out memfootprint_full_int_52184_pin.out ppm_full_int_52184_pin.out stride_full_int_52184pin.out On Tue, Oct 8, 2019 at 5:20 PM Amir H. Ashouri @_. wrote: … <#m-7586164980969694812> Not that I have known of. Do you see the output files when attaching MICA to it? Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations. - Amir Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ <http://www.eecg.toronto.edu/~aashouri/ > On Tue, Oct 8, 2019 at 11:14 AM dkokron @_.***> wrote: > I understand that MICA was primarily tested under pin-2.10. I have MICA > running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation > case I can use to test my installation? > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > < #16 <#16>?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA >, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q > > . > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#16 <#16>?email_source=notifications&email_token=ACODV2G4XN47NOZY2TSQ2GTQNUBS3A5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAVZQ7A#issuecomment-539727996>, or mute the thread https://github.com/notifications/unsubscribe-auth/ACODV2EEFCMKICURSY7OP2TQNUBS3ANCNFSM4I6TEA2Q . Hi dear @dkokron https://github.com/dkokron I'm using MICA these days and get some outputs in issue23, one typical example is hello.cpp, if you are interested in the outputs, we can have a detailed discussion about them :) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#16 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACODV2FCNGGCGQH4OD3IT2DTNFKGVANCNFSM4I6TEA2Q .

thank you @dkokron

Do you still remember how to compute ILP32 ? the output value of cpuClock_interval_all[0] in mica_ilp.cpp is 570000, but as far as I know, ILP 32 is less than 32.

dkokron commented 3 years ago

Sorry, no. I have no recollection.

On Wed, May 12, 2021 at 12:46 AM william @.***> wrote:

william, I haven't used MICA in quite some time and I have moved on to other work. Dan … <#m1304975651743258056> On Tue, May 11, 2021 at 11:18 AM william @.> wrote: Amir, I do get output in the various log files. I just don't want to trust the numbers without some validation. I'll try using pin-2.10 and see if I get similar results. ilp_full_int_52184_pin.out itypes_other_group_categories.txt memstackdist_full_int_52184_pin.out reg_full_int_52184_pin.out itypes_full_int_52184_pin.out memfootprint_full_int_52184_pin.out ppm_full_int_52184_pin.out stride_full_int_52184_pin.out On Tue, Oct 8, 2019 at 5:20 PM Amir H. Ashouri @. wrote: … <#m-7586164980969694812> Not that I have known of. Do you see the output files when attaching MICA to it? Alternatively, If you have the flexibility of installing it on Pin-2.10, you can run the same program/command multiple times, i.e., the same environment, isolated, etc. and consider an average of the outputs. Remember, this is a dynamic characterization technique, so even /bin/ls will output different values with minor fluctuations. - Amir Amir H. Ashouri Postdoctoral Fellow, Ph.D., P. Eng. University of Toronto, CA http://eecg.toronto.edu/~aashouri/ <http://www.eecg.toronto.edu/~aashouri/

On Tue, Oct 8, 2019 at 11:14 AM dkokron @.***> wrote: > I understand that MICA was primarily tested under pin-2.10. I have MICA > running under pin-3.11-97998-g7ecce2dac-gcc-linux. Is there a validation > case I can use to test my installation? > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > < #16 https://github.com/boegel/MICA/issues/16 <#16 https://github.com/boegel/MICA/issues/16>?email_source=notifications&email_token=ACECKHSEFDX7YJROCVA7OYTQNSPUHA5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HQMCJWA , > or mute the thread > < https://github.com/notifications/unsubscribe-auth/ACECKHTFTWUITYV3PM3X5JDQNSPUHANCNFSM4I6TEA2Q

. > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#16 https://github.com/boegel/MICA/issues/16 <#16 https://github.com/boegel/MICA/issues/16>?email_source=notifications&email_token=ACODV2G4XN47NOZY2TSQ2GTQNUBS3A5CNFSM4I6TEA22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAVZQ7A#issuecomment-539727996>, or mute the thread https://github.com/notifications/unsubscribe-auth/ACODV2EEFCMKICURSY7OP2TQNUBS3ANCNFSM4I6TEA2Q . Hi dear @dkokron https://github.com/dkokron https://github.com/dkokron I'm using MICA these days and get some outputs in issue23, one typical example is hello.cpp, if you are interested in the outputs, we can have a detailed discussion about them :) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#16 (comment) https://github.com/boegel/MICA/issues/16#issuecomment-838773316>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACODV2FCNGGCGQH4OD3IT2DTNFKGVANCNFSM4I6TEA2Q .

thank you @dkokron https://github.com/dkokron Do you still remember how to compute ILP32 ? the output value of cpuClock_interval_all[0] in mica_ilp.cpp is 570000, but as far as I know, ILP 32 is less than 32.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/boegel/MICA/issues/16#issuecomment-839466373, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACODV2HA4QEVBYHPACZUMMTTNII4ZANCNFSM4I6TEA2Q .